The Church of Betting

Premier League Draws

Value Betting Strategies – it has been a while since I have last written about the favourite topic of so many of you. It is not because I have lost interest in it – I continue to put small stakes on some of the strategies shared by Cassini in his Green-All-Over blog and occasionally test my own ones. Sometimes they work, sometimes they don’t. I just rarely feel that a potential trend (which might as well turn out to be just another manifestation of betting variance) is worth its own article. However, today the case is different. I would like to present you the Premier League Draws.

Soccermatics

A draw betting strategy has been making the rounds after having been covered by David Sumpter in his book Soccermatics: Mathematical Adventures in the Beautiful Game. The strategy resonated around the betting world and the book author has recently published an article about it on Pinnacle. So what exactly is this money maker? The strategy suggests backing the draw in games between equally matched teams in Premier League.

It turns out for that for years in all tiers of English football the market was pricing draws in games between equally matched teams too generously. In contrast, draws in games with large difference between the strengths of the opponents were overpriced. The distribution (from the article above) looked something like this:

Soccermatics Distribution

 

The founder of this pricing anomaly Mr Sumpter has decided to place some bets on it achieving a pretty decent return…

Soccermatics Profitability
Notice the 50% drawdown from game #200 till game #300 of this perfectly profitable strategy. Ouch!

…which convinced him to write a book and tell the world about it. An edge shared is an edge halved I hear you say. Surely, Mr Sumpter has been having second thoughts regarding his decision as he himself admits in the article. In fact, since the book has been published (May 5, 2016) the edge has not only been halved but has completely disappeared, as the article reports. Or has it really?

Analysis

I decided to check that for myself. Having collected the data from the great football-data.co.uk I have analysed the profitability of backing draws in all tiers of English football using Pinnacle closing prices since Season 2012/13 (the first one with Pinnacle closing prices on record). Here is the season-long profitability of a strategy of backing draws for all games in all leagues (blue line) compared to backing the draw where the difference between the winning probabilities of the two teams pre-game was less than 10%, 7% and 5% respectively:

English Draws / Soccermatics / All Leagues

The sample is indeed short and the returns volatile but one does see a downward trend for the betting strategy, whereas the blue line (all games) remains relatively flat. The premium of the strategy seems to have disappeared as we see the rest of the lines closing on and crossing the blue one in season 2017/18. That seems consistent with the findings of Mr Sumpter from his article and supports the intuitive notion that an edge so popular would have by now disappeared.

It makes sense to double check that by having a look at the opposite strategy, namely backing the draw in games with clear favourite and outsider. According to Soccermatics the draws for this subset were previously overpriced. I have compared the profitability of all draws with draws for games where the pre-game winning probability difference between the teams was at least 30%, 40% and 50%.

English Draws / Soccermatics / All Leagues / Contra

The results are consistent with the earlier ones in that the draws for games with clear favourites seem to have been underperforming all draws for some time. Later on the rest of the lines caught up with the blue one and in season 2017/2018 actually delivered a premium and a profit. It is perhaps too early to conclude that the trend has reversed but that is certainly worth further investigation.

How about the Premier League?

Before closing the case I decided to have a look at the results per league. After all the original strategy has focused on the Premier League in particular. I was surprised to find out that the Premier League still seems to deliver a premium for backing draws of evenly matched teams. The graph for backing draws between evenly matched teams for the Premier League alone looks like this:

Premier League Draws / Soccermatics / Premier League

Here again there seems to be a downward trend, however for most seasons the lines (excluding the blue one) remain in the profit zone. The premium compared to the blue line does not seem to have disappeared. Backing the draw for evenly matched teams in the Premier League continues to be a profitable strategy. Here are the numbers for backing the draw where there is less than 10% difference in winning probabilities (with level stakes):

Season Bets Implied Probability Actual Probability Total Return Edge Delta Edge (comp. to base case)
2017/18 24 31.04% 37.50% 5.46 6.46% 7.43%
2016/17 60 30.45% 36.67% 11.59 6.21% 8.97%
2015/16 79 29.47% 34.18% 13.27 4.70% 2.60%
2014/15 68 30.01% 32.35% 5.31 2.34% 3.62%
2013/14 52 29.74% 36.54% 12.51 6.80% 11.29%
2012/13 58 29.43% 39.66% 20.52 10.22% 7.31%

How does the graph look like for the opposite strategy – backing the draw for games with clear favourites in the Premier League?

English Draws / Soccermatics / PL / Contra

Not too optimistic, since the gap between the games with favourites and all games seems to have closed. Yet, compared to the graph for all leagues the trend at least has not reversed but at most evened up with the general case. This might be attributed to the larger initial gap as compared to the all-leagues-graph.

By the way, I have checked other leagues as well for the same trend and have discovered similar (although less pronounced) results for La Liga (again, level stakes, <10% difference in winning probability):

Season Bets Implied Probability Actual Probability Total Return Edge  Delta Edge (comp. to base case)
2017/18 28 29.93% 21.43% -8.01 -8.50% -7.37%
2016/17 64 30.21% 34.38% 9.31 4.17% 5.05%
2015/16 56 29.76% 30.36% 1.52 0.60% 0.82%
2014/15 67 30.05% 34.33% 9.62 4.28% 4.53%
2013/14 58 29.47% 34.48% 9.62 5.02% 6.35%
2012/13 56 28.61% 33.93% 10.09 5.32% 8.05%

Here too you will find steady premiums for backing the draw between equally matched teams as compared to all draws. The current year is a notable exception, however this can be attributed to the abnormally low frequency of draws in this bracket compared to previous seasons so reversion to the mean should be expected.

The analysis did also find out that in the second tiers of Spanish and Italian football as well as in Eredivisie backing the draw generally appeared to be a profitable strategy for the last 6 seasons. Serie B also shows some premium for backing the draw between evenly matched teams but it can mostly be attributed to the last season, which had too few such games so far so no conclusions can be drawn here.

To back or not to back?

To summarize, I do believe that backing the draw for Premier League games between equally matched teams remains a strategy with strong potential. Yet caution is warranted, since the counter-strategy is pretty much in line with the general draw profitability. Furthermore it doesn’t make intuitive sense that the bias has disappeared everywhere but in the Premier League. The Premier League is supposed to be the market with the highest volume and the smartest money in it so if anything you would expect the least market inefficiencies to be found there. And finally, we shouldn’t forget that the edge has been shared within a well-selling book so we shouldn’t expect whatever is left from it to be around much longer.

However, this surely continues to look like a profitable strategy so with the above concerns in mind I am currently putting small money on backing the draw on games where the difference of winning probability between the two teams is less than 10% in the Premier League. If you are interested in how the strategy is going just send me an email anytime during the season and I’ll let you know.


Deals

Some news from the betting services I’m partnering with:

rebel betting

RebelBetting has introduced a one-week trial for their arbitrage software. Try it out for free and if you are not happy with the product, just unsubscribe before the trial is over at no cost. There is not much to lose!

 

Furthermore, PremiumTradings is offering free deposits with Skrill, Neteller and Ecopayz for the month of December so you still have a few days left to make use of this offer.

 


 

Finally, I want to wish to all of you a Merry Christmas and a Happy New Year. 2017 has been a great year for me and I am looking back at it with a smile. A big Thank You to everyone for the nice mails and comments you have sent me throughout the year or even just for sticking around and reading my articles – this is what makes the experience of running The Church of Betting worth it. I hope you all have a great holiday season and see you soon in 2018!

 

5 thoughts on “Premier League Draws

  1. have a look at the time of the opening goal in an individual league in terms of expectation of a draw > would you have wanted to back the draw in the Liverpool game V Man City if you were aware that the last 5 completed seasons in the Premier League where the home team opened the scoring in the 0-10 time band and 1-0 with 11 minutes elapsed > sample size 199 and only 30/199 ended a draw = 15.08% = 6.63

    Instead of backing the draw before a game starts > have a look at backing / laying the draw in an individual league around the time of the opening goal added to the game state as time decays

    • Once an opening goal has been scored by either of two evenly matched teams, the teams are not evenly matched anymore, therefore at that stage the game is not anymore a qualifier according to the system. After an early goal not only are the odds turned in favour of the scoring team but also the probability of an over increases, which is bad news for draws.

      Alternatively, you could back the draw live in a game where the pre-game outsider scores a surprising opener and evens out the odds between the two teems. Unfortunately I have no live odds data available to see in how far that would have been a successful strategy but it would be in line with the philosophy of the strategy described in the article.

      Anyhow, if I wait for an opening goal in order to back the draw I miss the chance to back the draw for the games finishing 0-0, so in case there is any value in backing the draw in these (which I think there is) I would miss out on that value.

Leave a Comment